問題文
A logistics platform embeds a reinforcement learning component that keeps adjusting its routing policy in the field. The team runs its automated suite once before the release and then never again. Which practice does the syllabus call for with such a component?
選択肢
- Rerun the load suite whenever the traffic doubles, and watch whether the observed latency has grown beyond the ceilings that the operators tolerated.
- Rerun the smoke test whenever the vendor patches, and watch whether the installed version has moved beyond the baseline that the auditors passed.
- Rerun the whole suite whenever the platform reboots, and watch whether the recorded output has changed against the snapshot that the developers captured.
- Rerun the core suites whenever the components shift, and watch whether the measured levels has slipped beneath the threshold that the stakeholders accepted.